View RSS Feed

ratstsrub

Legion: Skin Deep and why storing information in DNA is awesome and feasible BUT NOT THAT WAY

Rate this Entry
3 Comments
by
ratstsrub
, April 27th, 2015 at 11:31 PM (4468 Views)
So I started reading Legion by Brandon Sanderson and all's fine and dandy until he reached I3, a company that will REVOLUTIONIZE information storage by transfecting human cells with bits of DNA to store information, using a virus.

And I'm just like, NO NO NO WHY WOULD YOU DO THAT.

For reference, the technology is totally feasible, DNA is just strings of 4 nucleotides, and as such, can totally be used to encode information. Rather like how we have 0 and 1 in transistors, now we have A, T, G, or C, with which we can convert to 0s and 1s and back quite easily since it's just a matter of encoding.

Read: http://www.nature.com/nature/journal...ture11875.html

So, it's a real thing, storing information in DNA. BUT, why the heck would you want that DNA to be in a cell?! Why would you want it to be part of your genome?

See, here's what's gonna happen when that DNA is in your genome, assuming that your delivery method is 100% safe and effective, it's gonna be subjected to a fuckload of damage. It's in a goddamn body, that's gonna be in the sun, that's gonna be injured, that's gonna be running fuck knows where, exposed to fuck knows what. Your bits of DNA is gonna be subjected to all that, and then you risk mutating it. Congrats, your information's now corrupted, hope you like that you cheeky little startup.

And that's just the external damage that's gonna risk mutating your bits of DNA that you stuck in your genome for no goddamn reason. Since, that doesn't even take into account the fact that, if you're wanting your DNA to be replicated (why the hell would you want that, just one molecule is enough and if you can make one molecule in the first place you can make thousands of it already!), there's an inherent mutation rate subjected to the accuracy of DNA polymerases, the little protein that copies your DNA, and fuck tons of other helper proteins and processes and whatnot. So, the more you copy your bits of DNA, the more likely it's gonna be mutated, and WHOOP THERE GOES YOUR DATA.

Just make the DNA molecules, with your information encoded on it, put it in a storage tube, then freeze dry that sucker and keep it in a cool and dark place and it'll last you thousands and thousands of years. There you fucking go, micrograms to milligrams of DNA storing your oh so important data with GB to TB of storage in less space than your goddamn thumb.


Urgh, now I can't even get through the rest of the story since this is BOTHERING ME TOO MUCH.


*Hell, even if you want to store it in cells, you already have the perfect host for such a thing, and the story itself even mentions it.

JUST STORE IT IN E.COLI CELLS. Yes! The bacteria where you don't give a damn if it has cancer because it can't get cancer! Or, fuck, I dunno, a plant, like corn.

Gah the more I think about this the more it irritates me.

Updated April 27th, 2015 at 11:34 PM by ratstsrub

Categories
Uncategorized

Comments

  1. RacingeR's Avatar
    I3
    I knew I3uster was evil.
  2. gesalt's Avatar
    Actually, a lot of human DNA is basically junk data so something like this would be fine. DNA is also fairly resilient though bits are shaved away every time a cell divides (negligible data loss over all except over a period of decades). Still not realistic but it would be neat to be able to access a ton of data by using some sort of reader that pricks you for a little blood.
  3. ratstsrub's Avatar
    RE: http://en.wikipedia.org/wiki/Jesse_Gelsinger on adverse reaction to viral vector based delivery method. In any case, you can't do it by random insertion, since the risk is still too high. Basically, would you take a 1 in 5 chance of it hitting something functional? Realistic speaking, your chance of getting cancer from that hit isn't all that high, but if it's gonna be a standard method of treatment, targeted insertion would be a lot better. In fact, there's a recent paper that just came out on this subject:

    http://link.springer.com/article/10....238-015-0153-5

    Where they used the CRISPR/Cas9 system (a bacterial DNA targeting and cutting system) on human embryos to replace a mutant Hemoglobin gene with a corrected copy (to treat Beta Thalassemia, I believe). The result was a mosaic embryo with some cells that had the correct replacement, some cells where the correct gene were inserted somewhere else, and some cells that just weren't affected at all. (BTW, the article is open access, so you can actually download it! Go ahead, facebook it for your friends to condemn men for playing gods )

    And this is the state of the art in terms of targeted gene replacement, so we're nowhere close to clinical applications.

    Anyways, on data corruption, it's not an issue of telomere lost, where the end bits are shaved away, but issue of just random base changes upon replication, that's the cause of data corruption over time. And we hates random.


    Now, ignoring all that, it would be totes neat to have a personal sequencer that would automatically a) isolate DNA from a given tissue sample, and b) sequence and give an accurate readout within an hour, that c) fits on a bench.

    Holy grail of sequencing technology right there. If you can make something like that, you'd be rolling in money. Everybody would want a piece of you. Personalized medicine, rapid diagnostic of bacterial infections, etc . . .
    Updated April 29th, 2015 at 12:46 AM by ratstsrub